This repository contains portfolio assignments (as a single pdf file), which were done during the experimental methods III course as part of Cognitive Science BA´s degree at Aarhus University
Teacher:Riccardo Fusaroli
Instructor: Kenneth Christian Enevoldsen
2019 fall semester
This course is about (some aspects of) how do you use statistical techniques in the real world (aka data science). It enables to master the complexity of the data: data juggling with tidyverse, generalized mixed effects linear models; timeseries analysis. Teaches how to create kick-ass generalizable and accountable models: Basic machine-learning techniques (explanation vs. prediction), model comparison, feature selection. Develops the skills necessary to build “data products”: shareable code repositories on github and meaningful analysis reports.
File Experimental_methods_3_Portfolio.pdf
contains a compilation of written reports solving the assignments listed below. Assignments code
folder contains rmd and md files displaying the code solving the assignment.
Assignment | Description |
---|---|
Portfolio 1: Data cleaning & wrangling (data preprocessing for Portfolio 2) | Testing out the git integration with RStudio. Cleaning and transforming several data sets into just one containing only the variables that are needed for the analysis. Warming up tidyverse skills (especially the sub-packages stringr and dplyr) |
Portfolio 2: Assessing linguistic development (in ASD) (code + report) | Developing a tool (mixed effects or multilevel models) to assess and predict language development in children |
Portfolio 3: Automated assessment of schizophrenia through voice analysis (code + report) | Building and assessing an automated tool to “diagnose” schizophrenia from voice. Creating a logistic regression model, which can be applied as a classifier. |
Portfolio 4: Interpersonal coordination of physiology (code + report) | Critically assessing the presence and mechanisms of interpersonal coordination. Preprocessing and modeling heart rate and respiration data |
Portfolio 5: Meta-analysis of voice patterns in schizophrenia (code + report) | Critically collectig and organizing previous findings in a meta-analysis to guide future studies |
Knowledge: After completing the course, students will have gained knowledge of:
- The conceptual assumptions underlying linear and non-linear statistical methods and the implications of choosing between such methods
- Strengths and weaknesses of non-linear methods
Skills: After completing the course, students will be able to:
- Motivate the choice between linear and non-linear statistical methods of analysis
- Design empirical studies with particular consideration of the strengths and weaknesses associated with the statistical methods used
- Carry out non-linear data analyses
Competences: After completing the course, students will be able to:
- Critically choose between, compare and substantiate the best qualified statistical tools for a given data set and research question
- Explain the results of linear and non-linear statistical analyses and their relevance to the research question